Predicting the Sulfur Precipitation Phenomena During the Production of Sour Natural Gas by Using an Artificial Neural Network

Abstract Deposition of elemental sulfur has been recognized as an important problem in the production of sour natural gas. This article presents a new approach using an artificial neural network (ANN) model for predicting the solubility of elemental sulfur in reservoir sour gases of various compositions. The proposed three-layer feed-forward neural network model is much more accurate than the phase equilibrium model for predicting the solubility of sulfur in super/near-critical sour natural gas mixtures at reservoir and well tube operating conditions and is reliable for evaluating the risk of sulfur precipitation during sour gas production.